On statistical investigation of the virtual water trade network

A modeling approach for studying the networks of virtual water flows is proposed here. Particularly, the mixed effects model (i.e. random effects models and their generalization) approach has been chosen, also for its advantage concerning availability of good estimation methods. In the network of virtual water trade each country participating in trade is represented by a node. Links among nodes are the volume of virtual water embodied in the traded commodities and are directed on the base of direction of trade flow. Mixed effects models for network data are practical to implement also in the case of valued networks, such as those where counts, measurements, or ranks are observed. In the application, detailed international trade data from FAOSTAT, to reconstruct the global trade patterns of each food product reported by the FAOSTAT data base, including crops, crop-derived food commodities, and animal products, are used. Then, the country specific average water footprint of each product by crop FAOSTAT code, for the conversion of the crop trade matrices to virtual water trade matrices, is used. The results of the preliminary analysis are not presented here for the sake of brevity but more investigations are encouraged. Above all, as in the mentioned database, the virtual water content of many products is available also at sub-national scale, investigation on the domestic virtual water trades and its management are also foreseen. Understanding the virtual water trade concept and strategy may be important for formulating informed policies and improving water use efficiency at different levels, i.e. for those countries relatively dry in some areas and relatively wet in other areas, domestic virtual water trade may be a relevant issue.